Skip to main content
Hello Dr. Quiniou,
Thank you for your question on our results, I'm happy to clarify. The main data during our intervention period came from a report pertaining to the intervention BPA (best practice alert) in the EHR. This report contained information for all encounters which the BPA fired with the corresponding response to that BPA and additional patient and encounter details. We aggregated this data ourselves to the individual patient level which is where we had the 14,411 patients.
To evaluate whether the number of referrals had increased we had queried the EHR the year prior to the start of the BPA to capture the number of patients similar in nature (I.e.meeting the same criteria which would have fired the BPA) seen at our intervention sites and whether or not they had received a referral in that baseline/pre-intervention time period. That corresponds to the 19,160 patients at baseline. To ensure that our calculation of intervention time period rates didn't deviate purely because of the way we captured the data, we also repeated this EHR query for the intervention time period which resulted in the 24,885 patients. Thus, there were two sources of data for the intervention period (the BPA report and the EHR query) which were both compared with the EHR query at baseline. The reason the 14,411 patients and 24,885 patients for the two intervention data sources do not match up is due to several factors, the largest being that at one site we could only use patients who signed and completed the HHI and formal consent to use the individual level BPA encounter information. The EHR query data was provided to us in de-identified aggregate form and so we were able to obtain information on more patients.